Introduction The evidence base regarding characteristics of older adults with bipolar disorder (BD) remains limited. The NIH‐funded multicenter study Acute Pharmacotherapy of Late‐Life Mania (GERI‐BD) assessed various clinical domains before and during mood stabilizer treatment in older adults participating in a 9‐week, double‐blind randomized controlled trial. We describe the rationale for selecting these instruments. Methods Domains and instruments were selected on the basis of the study design and the participants. The investigators' experience in clinical trials involving young adults with BD or older adults with major depressive disorder, along with open studies of older adults with BD, contributed to the selection process. Results We identified domains and selected instruments that could be used to assess the participants given their diagnostic, treatment history, and medical and mood state characteristics. They were also intended to measure tolerability and efficacy and permit examination of potential moderating and mediating factors. Conclusions Decisions regarding the assessment domains to be included in the clinical trial highlight the challenges facing researchers studying drug treatments for older adults with BD, or more generally, mood disorders. We suggest that the domains and instruments selected by GERI‐BD investigators constitute a “toolbox” that can be customized for other investigators. Copyright © 2017 John Wiley & Sons, Ltd.